Game Day 12 – The Rememberance Day weekend.
by Gordon Haverland
Largely in rememberance of WWI and WWII. Matic has memories of a Balkin war that bothers him.
Game Day 12 takes place over 2 physical days (Saturday and Sunday). On Saturday there is one early game, 4 games running in the middle time slot, and one late game. On Sunday, there is an early game, a not quite so early game, and then 2 late games (probably trying to cut the Arsenal audience down, why else play that at the same time as Man$ity v ManU?).
Much of what follows is written to a “script”, and there can be fine points that are not covered.
In what follows we have an analysis of how many times each team has inflicted the need for treatment on the others. So you can see where this is going here’s a summary – the full details of the effect of the treatment is shown in the text. All figures are just for this season. More analysis of the data will follow later.
- Everton have inflicted the need for treatment on opposition players 17 times
- Tottenham have inflicted the need for treatment on opposition players 16 times
- Cardiff players have inflicted the need for treatment on opposition players 13 times
- CPalace have inflicted the need for treatment on opposition players 13 times
- Southampton inflicted the need for treatment 13 times
- Fulham have inflicted the need for treatment on opposition players 13 times this season
- Huddersfield players have inflicted the need for treatment 11 times
- Newcastle has inflicted the need for treatment on the opposition 11 times
- Arsenal have inflicted the need for treatment on the opposition 11 times
- Bournemouth has inflicted the need for treatment on the opposition 11 times
- West Ham (StateAid) players have inflicted the need for treatment 11 times
- Liverpool! have inflicted the need for treatment on opposition players 10 times
- Watford inflicted the need for treatment on opposition 10 times
- Burnley players have inflicted the need for treatment 8 times
- Chelsea have inflicted the need for treatment on opposition players 8 times
- Brighton players have inflicted treatment 6 times
- Leicester players have inflicted the need for treatment 6 times
- Wolverhampton have inflicted the need for treatment on opposition players 5 times
Here’s a look at the individual matches
Cardiff – Brighton
Referee: Martin Atkinson
Assistants: Stephen Child, Peter Kirkup
Fourth official: David Coote
With new fangled data, I see that Cardiff has 596 man-minutes of caution so far this season (7th) and Brighton has 856 (19th).
Cardiff players have needed treatment 16 times this season, required a substituion on 5 of those occassions and had to play short-handed for 36 minutes (a kind of caution). Brighton players have needed treatment 16 times this season, required a substitution on 6 of those occassions and had to play short handed for 36 minutes.
Cardiff players have inflicted treatment 13 times, requiring the opposition to substitute players on 2 occassions, causing the opposition to play short handed for 17 minutes. Brighton players have inflicted treatment 6 times not requiring any substitutions and only causing the opposition to play short handed for 13 minutes.
Martin Atkinson has managed 19 treatments this season (9 to home). The amount of time teams have to play short-handed due to his management is 21 to home and 22 to away. The away team had to substitute 4 times, and the home team 3 times.
Huddersfield – StateAid
Referee: Chris Kavanagh
Assistants: Simon Bennett, Marc Perry
Fourth official: Michael Oliver
Huddersfield has played under 656 man-minutes of caution (10th) and StateAid has played under 668 man-minutes (12th).
Huddersfield players have needed treatment 14 times, requiring 1 substitution and playing short-handed for 24 minutes. State Aid players have also required treatment 14 times this season, requiring 4 substitutions. They had to play short-handed for 33 minutes.
Huddersfield players have inflicted the need for treatment 11 times this year, requiring the opposition to substitute players 5 times. In the process, their opposition has played short-handed for 12 minutes. StateAid players have inflicted the need for treatment 11 times, and causing the opposition to need a substitute twice. Their opposition had to play short-handed for 34 minutes.
Chris Kavanagh has managed 13 treatments (7 to home teams). The amount of time teams have to play short-handed is 11 to home and 10 to away. Three home teams needed to substitute players and 1 away team.
Leicester – Burnley
Referee: Mike Dean
Assistants: Darren Cann, Dan Robathan
Fourth official: Stuart Attwell
Leicester has played under 798 man-minutes (15th) of caution, and Burnley has played under 740 man-minutes (14th).
Leicester players have needed treatment 14 times this year, requiring 2 substitutions and played short-handed for 17 minutes. Burnley players have needed treatment 11 times and required substitution 3 times, requiring them to play short-handed for 14 minutes.
Leicester players have inflicted the need for treatment 6 times, none of which required a substitution. Their opposition had to play short-handed for 11 minutes as a result. Burnley players have inflicted the need for treatment 8 times, requiring the opposition to substitute players twice. Their opposition had to play short-handed for 37 mintues as a result of this.
Mike Dean has managed 14 treatments this season (9 to the home team). The amount of time teams have had play short handed is 12 to home and 11 to away. Both home and away had to substitute players twice.
Newcastle – Bournemouth
Referee: Lee Probert
Assistants: Simon Beck, Andy Garratt
Fourth official: Craig Pawson
Newcastle has played under 421 man-minutes of caution (3rd) and Bournemouth has played under 590 man-minutes (6th) of caution.
Newcastle players have required treatment 14 times this season, needing to substitute players twice. In the process they had to play short-handed for 21 minutes. Bournemouth players have required treatment 14 times this season, needing a substitute on a single occassion. In the process, they had to play short-handed for 21 minutes.
Newcastle has inflicted the need for treatment on the opposition 11 times, causing the opposition to substitute players 5 times. Their opposition had to play short-handed for 23 minutes as a result. Bournemouth has inflicted the need for treatment on the opposition 11 times this season, causing their opposition to substitute players twice. In the process, their opposition had to play short-handed for 8 minutes.
Lee Probert has managed 17 treatments this year, 7 to home teams. The away team needed to substitute players 3 times, and home never did. Both home and away played short-handed 10 minutes due to treatments.
Southampton – Watford
Referee: Simon Hooper
Assistants: Richard West, Harry Lennard
Fourth official: Andre Marriner
Southampton has played under 680 man-minutes of caution (13th) and Watford has played under 804 man-minutes of caution this season (16th).
Southampton players have required treatment 18 times this season, and needed to substitute players on 4 occassions. They had to play short-handed for 30 minutes due to treatments. Watford players have required treatments on 3 occassions, and needed no substitutes. They had to play short-handed for 9 minutes for those treatments.
Southampton inflicted the need for treatment 13 times, requiring opposition to substitute 3 players. In that process they had to play short-handed for 16 minutes. Watford inflicted the need for treatment on opposition 10 times this season, requiring opposition to substitute players 3 times. In the process their opposition had to play short handed for 18 minutes.
Simon Hooper has managed 1 treatment this season (to away). No substitutions were required, and that away team played short-handed for 2 minutes.
CPalace – Spuds
Referee: Jonathan Moss
Assistants: Eddie Smart, Andy Halliday
Fourth official: Roger East
CPalace has played under 662 (11th) man-minutes of caution and the Spuds have played under 481 man-minutes of caution (5th).
If you do not trust the PGMO officials to behave in an unbiased manner, card and foul data are useless as a measure of how dirty a team is. At the moment the spuds are tied for first in one measure of dirtyness and in second place in another measure. And yet, they only have 481 man-minutes of caution applied to them. Odd.
CPalace players have needed treatment only 4 times this season, and not been required to substitute players as a result. They have only needed to play short-handed for 7 minutes as a result of this. Spud players have required treatment 10 times this season, and needed to substitute players on 5 occassions. They have had to play short-handed for 23 minutes as a result of this.
CPalace have inflicted the need for treatment on opposition players 13 times, which resulted in the opposition substituting players 3 times. In that process, their opposition had to play short-handed for 23 minutes.
The Spuds have inflicted the need for treatment on opposition players 16 times, which required opposition teams to substitute players 5 times. In the process, spud opposition had to play short-handed for 21 minutes.
Jonathon Moss has managed 17 treatments this season, 6 to home teams. No home teams needed to substitute players, but 2 away teams did. Away teams were forced to play short-handed for 21 minutes, while home teams only had 9 minutes short-handed.
Liverpool! – Fulham
Referee: Paul Tierney
Assistants: Adrian Holmes, Mick McDonough
Fourth official: Chris Kavanagh
Liverpool! has played under 347 man-minutes of caution (1st) while Fulham has played under 972 man-minutes of caution (20th).
Liverpool! players have needed treatment 11 times this season, requiring 5 players to be substituted. In the process, the played short-handed for 14 minutes. Fulham players have needed treatment 12 times this season, requiring 4 players to be substituted. In the process, they have played short-handed for 24 minutes.
Liverpool! players have inflicted the need for treatment on opposition players 10 times this season, resulted in one substitution of an opposition player and having them play short handed for 15 minutes.
Fulham players have inflicted the need for treatment 13 times, resulting in one opposition substitution. In the process, Fulham opposition have played short-handed for 23 minutes.
Paul Tierney has managed 11 treatments this season (5 to home teams). One home team required a substitution, no away teams did. Only 11 minutes of playing short handed resulted (5 to home).
Chel$ea – Everton
Referee: Kevin Friend
Assistants: Lee Betts, Scott Ledger
Fourth official: Andre Marriner
Chel$ea lies 2nd in the caution league with 388 man-minutes of caution. Everton lies 17th at 814 man-minutes of caution.
Chel$ea players have needed treatment 5 times this season, and needed no substitutions. They have played short handed for 6 minutes as a result. Everton players have needed treatments 14 times this season, requiring 4 substitutions. In the process they played short handed for 14 minutes.
Chel$ea has inflicted treatments on the opposition 8 times this season, which resulted in 1 substitution. Their opposition played short handed for 10 minutes. Everton has inflicted treatments on opposition 17 times, requiring 3 substitutions. In the process, their opposition played short handed for 25 minutes.
Kevin Friend has managed 12 treatments, split equally between home and away. Two home treatment required substitutions, while none for away. Home teams played short-handed for 17 minutes, while away teams played short handed for 5 minutes.
Arsenal – Wolves
Referee: Stuart Attwell
Assistants: Constantine Hatzidakis, Derek Eaton
Fourth official: Roger East
Arsenal lies in 9th place in the cautiod league, with 655 man-minutes of caution. We get cards for such horrible things as having garlic for lunch and breathing on people. It’s in the fine print in the FIFA Laws of the Game, and who reads the fine print? Wolves lie in 4th place in the caution league, with 433 man-minutes of caution.
Arsenal players have needed treatment 11 times this season, and this resulted in 2 substitutions and playing short-handed for 20 minutes. Wolves players have needed treatment 6 times this season for no substitutions and 15 minutes of being short-handed.
Arsenal has inflicted treatments on opposition 11 times for 1 substitution and 20 minutes of playing short-handed. Wolves have inflicted treatments 5 times requiring 2 substitutions and 16 minutes of being short-handed.
Stuart Attwell has manged 10 treatments (4 to home). Only 1 substitution to an away team. Away teams played shorthanded for 16 minutes, home for 3.
Man$ity – ManU
Referee: Anthony Taylor
Assistants: Gary Beswick, Adam Nunn
Fourth official: Craig Pawson
Man$ity lies 8th in the caution league, with 638 man-minutes of caution. ManU lies 18th in the caution league, with 828 man-minutes of caution.
Man$ity players have needed 7 treatments this season with 0 substitutions and playing short-handed for 11 minutes. ManU players have needed 5 treatments this season, for 0 substitutions and 2 minutes of playing short-handed.
Man$ity players have inflicted the need for 10 treatments on the opposition requiring 3 substitutions and playing short-handed for 14 minutes. ManU players have inflicted the need for 13 treatments and 4 substitutions and playing short handed for 21 minutes.
Anthony Taylor has managed 21 treatments this season (9 to home teams). There were 3 substitutions needed (1 to home). The home teams played short handed for 16 minutes and the away teams for 17 minutes.
Gordon,
Thanks for doing a preview with a different take on things for this weekends matches.
The only thing I will add is the appointment of Anthony Taylor to the Manchester Derby. Normally even the PGMP avoid putting local referees in charge of games, but here Taylor is a resident of Greayer Manchester living, I believe in Withenshaw which is a staunchly United supporting area.. this has angered a large section of the City supporters who are questioning his ability to deliver an impartial performance. As Arsenal supporters many of us question his competence as well but he retains the full trust of his masters at the PGMO. We should all watch that performance with interest to see how he does.
You’re welcome Andrew.
Even though I have all the data “here”, it still took quite a while to work it all up.
Now to see how to use R from within Perl. First on possession data, and then maybe on the foul and card data in conjunction with the possession data.
At the moment, I am looking at the possession data. The particular form of possession data I am looking at, is probably measured as the percent of time the home team has possession, and the away team is then assumed to be 100-P.
This data is only given to the nearest 1 percent, and ranges from about 20 to 80. There are (about) 60 unique values seen, so when one is working with more than 60 data points, there will be multiple incidences of some values. A plot of the cumulative distribution will show a stair-step effect from this. Since my data is complete up to the end of Game Day 11, I have 110 games of data, or 220 values of possession.
One thing I am trying to learn how to do, is to split this data into home possession versus away possession. For the entire data set, the number of home values will equal the number of away values. I will then use a non-parametric analysis of distribution to examine the difference between home possession and away possession. (I could plot these graphically, and “observe” the difference.) There are different tests that compare the two curves in different manners, and many of these tests allow you to say how likely the two curves are to be from the same distribution.
In football, one sort of expects the two distributions to be different.
The EPL seems to have two groups of teams:
1. Top-6
2. RestOfThePack (ROTP)
The Top-6 consists of 6 teams, the ROTP consists of 14 teams. So, to compare Top-6 and ROTP, the two distributions have different amounts of data. This complicates things.
I sort of expect Top-6 to be different from ROTP.
If it seems that home and away are different, it makes sense to look at home and away aspects of the Top-6 and ROTP.
As the Top-6 play ROTP (and each other), we have 4 categories of games:
1. Top-6 v Top-6
2. Top-6 v ROTP
3. ROTP v Top-6
4. ROTP v ROTP
And so on. Every time we add complexity, we divide up our pool of 220 data points into smaller and smaller sets. At some point, there won’t be enough data to answer questions (now). We are approaching 1/3 of the season, and a person may be able to guess at what a person might be able to do come the end of the season.
One of the early tests for distributions is the Kolmogorov–Smirnov test (K–S test or KS test).
https://en.wikipedia.org/wiki/Kolmogorov%E2%80%93Smirnov_test
There are newer tests which seek to avoid some problems with the KS test.
Some trivia (no analysis).
In mixed games so far this season, Arsenal have played 4 games at home and 4 games away (only Top-6 like that).
Chelsea and ManU have played 5 mixed games away, and the spuds have played 6.
Man$ity have played 5 mixed games at home. And the spuds are avoiding home games in the early season, have only played 2 mixed games at home.
The R-Project kSamples module has many functions in it, and they are all new to me.
I am starting with the Anderson-Darling test, but I am not sure if it is even appropriate, so the results may be meaningless.
Comparing the home possession data to the away possession data, this test thinks those two distributions are very different. There are 110 data points in each of those vectors.
Comparing the Top-6 home possession to the Top-6 away possession, the test thinks these are different (only 9 data points in these 2 vectors).
Comparing Rest-Of-The-Pack home and away possession, this has a better chance that they are different. In large part, it is because there are 53 data points in each of those vectors.
I then looked into the home possession behavior of Top-6 versus ROTP. This is a vector of 9 points and a vector of 53 points. It is likely this test has no power here.
The above isn’t an analysis, it is just learning how to use the tools.
I think I am going to have to put some work with this on hold. I have other things which need doing.
But, like with any “population” of people, statisticians have their own language. And nobody provides good enough documentation. So, I need to build a test system where I generate synthetic data of known properties, and then run things against the functions in kSamples, to figure out how to translate their answers into something I understand.
I will continue to add GameDay data to what I have now, to work with data similar to what was in this thread article. But, as so few people commented, I don’t think anyone is interested.